Sep 2025 – Present · 6 mos

Dec 2016 – Sep 2025 · 8 yrs 9 mos

• 1️⃣ Designed & Developed End-to-End Reminder Creation Pipeline using MultiTurn Conversation Platform
• Built an end-to-end conversational pipeline that enabled back-and-forth voice interactions instead of single-shot commands.
• Allowed users to create, modify, and confirm reminders using natural, human-like conversations.
• Example user scenarios:
• User: “Remind me to call John tomorrow.”
• Cortana: “Sure, at what time should I remind you?”
• User: “10 AM.”
• Cortana: “Got it! I’ll remind you to call John tomorrow at 10 AM.”
• This multiturn capability significantly improved usability by making interactions intuitive and human-like.
• Scenario #2) Integrated Time-Based & Location-Based Reminder Scenarios
• Time-based reminders: e.g., “Remind me to join the meeting at 5 PM today.”
• Location-based reminders: e.g., “Remind me to send the report when I reach work.”
• 2️⃣ Built Data-Driven Insights to Improve User Experience
• => Analyzed production logs to understand user engagement patterns, latency, and performance metrics.
• => Derived actionable insights to identify
• 1) Where reminders were performing well ✅
• 2) Areas needing optimization ⚡
• 3) Leveraged these insights to recommend and implement other feature enhancements
• 4) performance improvements on reliability, speed, and usability.
• 3️⃣ Ensured High Availability Through Live-Site Monitoring
• => Led live-site monitoring initiatives to ensure the service remained available 24/7.
• => Developed and configured automated monitoring systems and alert mechanisms to proactively detect and resolve issues, minimizing downtime and maintaining a seamless user experience.

Jun 2015 – Nov 2016 · 1 yr 5 mos

• Bing Search Engine , Reminder Services On Cortana, Windows Client
• 1) Design and developed the pipeline E2E to create reminders that helps users. It helps users to remind certain actions , things at certain time , certain location and contact based as and when it is required .
• For example : - User can query "remind me to meet XYZ at 5 pm today" , "remind me to call my dad after i reach home" in cortana, windows client . Our service and scenarios reminds the users through notification ,other mechanisms "to meet XYZ at 5 pm that day" - a time based reminder and notifies to "to call my dad" after i reach home - location based reminder.
• 2) Worked on multi turn contextual platform for the creation of certain scenarios like reminders , uber - It understands the previous turns query from the users and responds accordingly for the queries in the next turn
• Bing Search Engine, Cortana
• 1) Design and developed the backend pipeline to show his/her Movie, Events, Restaurant
• Bookings Proactively On Windows 10 Cortana Canvas,
• 2) Developed Scenarios On Integration of Uber Cab Booking With Movie, Event, Restaurant
• Reservation Scenarios.
• If the User Has Movie Booking @ 1'0 Clock, Then Windows 10 Cortana
• Canvas shows both Movie Reservation & Uber Cab Booking Features From T-2 Hours
• i.e., from 11'0 Clock . It also allows him/her to book the cab to Movie from Cortana Canvas.
• 3) Developed the Pipeline to Retrieve Appropriate Data/JsonFrom Bing
• Search & Local Search REST API for the Cortana Scenarios.

Jul 2010 – May 2015 · 4 yrs 10 mos

• Bing Health & Fitness Segment .
• 1) Designed & Developed The Bing Answers For Health and Fitness Queries like "Calories In Banana", "Yoga For Weight Loss".
• 2) Designed & Developed The Entire Back end Automation Test Frame Work to Validate the Health And Fitness Answer Pipeline Starting Data Layer to UI Layer. This Has Been Incorporated By Many Teams .
• 3) Designed the Process To Validate The Health & Fitness Classifier to Ensure That Precision and Recall Measurements are good.
• 4) Automated Tests Of Health & Fitness iOS App Using Instruments Framework.
• Bing Weather Segment .
• 1) Designed & Developed Bing Answers For The Weather Queries like "Weather In Hyderabad".
• 2) Designed & Developed Live Site Monitoring Tool to Download Production Data Automatically to Repro The Issue in local box For The Faster Investigation & Mitigation.
• 3) Designed the Alerting Mechanism For Weather Answer - This Alerting Mechanism Helps Us To Know The Issues In Production Proactively in turn helped the team To Mitigate Before It Becomes a Live Site.
• 4) Designed And Developed The Data Manipulation Tool for the Weather Answer Data Validation.
• 5) Implemented a Methodology To Measure Latencies, Load, Performance Analysis of Weather Answer.